Synthesis of chiral N-protected 1,2-dihydro-quinoline-2-carbonitrile and 1,2-dihydro-isoquinoline-1-carbonitrile via an asymmetric Reissert reaction

Article abstract of DOI:10.1016/j.tetlet.2005.03.034

Addition of cyanide ion to chiral N-acyl-quinolinium and N-acyl-isoquinolinium salts led selectively to 1,2-addition products. Removal of the chiral auxiliary affords the title compounds in pure enantiomeric form.

Full text of DOI:10.1016/j.tetlet.2005.03.034

Tetrahedron
Letters
Tetrahedron Letters 46 (2005) 2983–2987
Synthesis of chiral N-protected 1,2-dihydro-quinoline-2-carbonitrile
and 1,2-dihydro-isoquinoline-1-carbonitrile via an asymmetric
Reissert reaction
Mickael Pauvert,a, Sylvain C. Collet,a Marie-Jo Bertrand,a
¨
a,
Andre Y. Guingant and Michel Evain
b,à
*
´
a
` ´
Laboratoire de Synthese Organique, UMR CNRS 6513, FR CNRS, 2465 Faculte des Sciences et des Techniques,
2, rue de la Houssiniere, BP 92208-44322 Nantes Cedex 03, France
`
Institut des Materiaux Jean Rouxel, 2, rue de la Houssiniere, BP 92208-44322 Nantes Cedex 03, France
b
´
`
Received 7 January 2005; revised 25 February 2005; accepted 7 March 2005
Available online 22 March 2005
Abstract—Addition of cyanide ion to chiral N-acyl-quinolinium and N-acyl-isoquinolinium salts led selectively to 1,2-addition
products. Removal of the chiral auxiliary affords the title compounds in pure enantiomeric form.
ꢀ 2005 Elsevier Ltd. All rights reserved.
The 1,2 addition of cyanide ion to N-benzoyl quinolin-
ium chloride generated in situ by the action of benzoyl
chloride on quinoline is known as the Reissert reaction.1
Other aromatic acid chlorides and nitrogen heterocycles
may be used to give the so-called Reissert compounds as
depicted in general terms in Scheme 1.
philes other than the cyanide ion could be added to
the intermediate iminium species.3 In conjunction with
several synthetic projects we became interested in
the use of the Reissert reaction to prepare differently
substituted 1,2,3,4-tetrahydro-quinoline derivatives.
Somewhat surprisingly, the problem of achieving an
asymmetric version of the Reissert reaction has not been
addressed until quite recently. The first proposed solu-
tion was disclosed by Shibasaki and co-workers4 who
discovered that a bifunctional Lewis acid embodying
the 2,20-binaphthol (BINOL) motif was able to catalyse
the cyanide addition to variously substituted quinolines
(isoquinoline) to give the corresponding Reissert
compounds in enantiomeric excesses ranging from
54% to 96%.
The synthetic potentialities of this venerable reaction2
have found wide applications in heterocyclic chemistry,
particularly since it was recognised that several nucleo-
CN
CN
N
Ar
N
N
Ar
ArCOCl
Following this pioneering work, Liebscher and co-work-
ers5 showed that the combined action of TMSCN
and (ꢀ)-(R)-menthylchloroformate onto isoquinoline,
in the presence of aluminium chloride, delivered the cor-
responding Reissert compound in greater than 95% dia-
stereoselectivity. The (S)-N-cbz-alanoyl fluoride was
also shown to be equally efficient as chiral inducer. Since
the Shibasaki protocol failed to give a high selectivity
with isoquinoline, the Liebscher approach complements
well the work of the Japanese team. These results
prompted us to disclose our own approach to the prob-
lem which makes use of an Evans chiral auxiliary to
Cl
O
Cl
O
Scheme 1.
Keywords: Asymmetric Reissert reaction; Chiral oxazolidinone-derived
auxiliary; N-protected 1,2-dihydro-quinoline-2-carbonitrile; N-pro-
tected 1,2-dihydro-isoquinoline-1-carbonitrile.
*
Corresponding author. Tel.: +33 02 51 12 54 13; fax: +33 02 51 12 54
02; e-mail: andre.guingant@chimie.univ-nantes.fr
`
Present address: Atlanchim company, 2 rue de la Houssiniere, BP
92208-44322 Nantes Cedex 3.
à To whom questions regarding X-ray structure should be addressed.
0040-4039/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2005.03.034

2984
M. Pauvert et al. / Tetrahedron Letters 46 (2005) 2983–2987
CN
CN
O
O
O
O
O
O
H
+
N
N
N
Cl
N
N
N
H
H
H
O
O
O
Cl
Ph
Ph
Ph
major diastereomer
formed?
2
1
Scheme 2.
Table 1. Reaction of 1a and 1b with TMSCNa
reach quinoline- and isoquinoline-derived Reissert
compounds with useful diastereoselectivities. We thus
envisioned that the bias usually provided by the chiral
4-benzyl-oxazolidin-2-one auxiliary would selectively
direct the cyanide anion on one particular diastereo-
topic face of N-acyl (iso)quinolinium chlorides 1 as
shown in Scheme 2. The salts 1, key intermediates
to this approach, would be in situ generated by
simply reacting (iso)quinoline with the known6 acid
chloride 2.
Entry Substrate Xꢀ
T (ꢁC)
Products
(Diastereomeric Yield (%)c
ratio)b
1
2
3
4
5
6
1a
1b
1a
1b
1b
1a
Clꢀ
Clꢀ
TfOꢀ 20
TfOꢀ 20
TfOꢀ 40
NfOꢀ 40
20
20
3:4 = 55:45
5:6 = 60:40
3:4 = 83:17
5:6 = 15:85
5:6 = 29:71
3:4 = 84:16
88
76
70
35
95
60
a All reactions were performed in methylene chloride.
To begin to investigate the behaviour of N-acyl salts 1
with cyanide anion, acid chloride 2 was prepared by
reacting the sodium salt of (4S)-4-benzyl-oxazolidin-2-
one with a toluene solution of phosgene (1 equiv) in
conditions previously reported.6 Reaction of 1a (1b),
generated by addition of 2 to a solution of quinoline
(Q) or isoquinoline (isoQ) in methylene chloride, with
TMSCN as a cyanide source, led to the expected form-
ation of diastereomeric Reissert compounds 3 (5) and 4
(6) as depicted in Scheme 3. The reaction was quite dis-
appointing in terms of selectivity since an almost equal
mixture of diastereomers were formed in each case
(Table 1, entries 1 and 2). In a second experiment, hop-
ing that the nature of the associated anion would
favourably influence the diastereomeric ratio of the
nitrile products, exchange of Clꢀ for TfOꢀ was effected
prior to the addition of TMSCN. As shown in Table 1
(entries 3 and 4) this operation significantly improved
the diastereomeric ratios which increased from ca. 1:1
to ca. 4:1 in both examples.7 Raising the temperature
b Determined by HPLC analysis (column: Inersil Si Chrompack,
3 · 250 mm; k = 254 nm; eluent: CH2Cl2; flow rate: 0.5 mL/min).
c Isolated combined yields after column flash chromatography.
to 40 ꢁC substantially improved the yield of the cyanide
addition to isoquinolinium ion although, in that case,
the ratio of diastereomeric adducts 5 and 6 was signifi-
cantly lowered (compare entries 4 and 5). Finally,
changing the triflate counterion for the more volumi-
nous nonaflate ion did not affect the diastereomeric ratio
of adducts 3 and 4 (entry 6) suggesting that the benefi-
cial role of triflate ion over chloride ion in controlling
the course of the reaction finds its origin in electronic
rather than in steric effects. Although our results are
inferior to those observed in the recent literature it
should be pointed out that no elaborated additive is re-
quired and that the diastereomeric adducts 3 (5) and 4
(6) are easily separated by standard silica gel flash chro-
O
O
N
Cl
2
2
H
isoQ
Q
O
N
O
N
Ph
X
X
R*
R*
O
R*
1a
1b
2
+
+
N
CN R*
5
O
N
O
N
CN
R*
N
CN
R*
CN R*
O
O
3
4
6
Scheme 3.

M. Pauvert et al. / Tetrahedron Letters 46 (2005) 2983–2987
2985
matography, rendering the reaction synthetically useful.
Moreover, chiral auxiliary removal can be easily per-
formed in one single operation (vide infra).
N+–CO bond. Attack of cyanide ion from the less con-
gested faces of conformations (C1) of salts 1a and 1b
should thus account for the preferential formation of
adducts 3 and 6, respectively. It is not without interest
to remark that the conformations displayed by these ad-
ducts in the crystalline state structures13 are closed from
conformations (C2). Finally, the importance of the asso-
ciated triflate anion on the diastereoselectivity of the
addition process could be that, being less tightly bound
to the iminium moiety than the chloride anion, it
favours the establishment of the cation–P complex.
Stereochemical assignment was unambiguously accom-
plished by subjecting each major diastereomer (entries
3 and 4 in Table 1) to X-ray crystal structure analysis.
The analysis revealed that these adducts correspond to
structures 38 and 6 (Fig. 1),9 respectively.
The reason for the above stereochemical trends is not
clear. A possible pathway that could be envisaged as
the major mechanistic possibility leading to adducts 3
and 6 would involve the formation of a stabilising cat-
ion–P complex intermediate.12 This in turn would then
lead to an effective screening of one of the diastereotopic
faces of the reacting C@N+ double bond. Considering
the two potential reactive conformations (C1) and
(C2) for iminium salts 1a and 1b, rotations around each
of the N–CO bonds are necessary to allow these salts to
escape to severe steric interactions (Scheme 4). As a re-
sult it became apparent from examination of a model
that the establishment of a cation–P complex is dis-
favoured in conformations 1a (C2) and 1b (C2) whereas
such a stabilising disposition remains easily possible in
conformations 1a (C1) and 1b (C1) having the iminium
and carbonyl functionalities trans-disposed to the single
We next focussed our attention on the chiral auxiliary
removal operation. To expand the utility of the above
reactions for the synthesis of 1,2,3,4-tetrahydro-
(iso)quinoline derivatives it appeared useful to remove
the chiral auxiliary while, in the same time, protecting
the nitrogen atom.
This could be simply and efficiently realised by
treatment of adducts 5 and 6 with samarium tri-
flate (0.2 equiv)14 in a mixture of methylene chloride
and methanol at room temperature. Following this
procedure, the reaction proceeded cleanly without com-
promising the chiral centre and led to the 1-cyano-1H-
isoquinoline-2-carboxylic acid methyl esters (R)-7 and
(S)-7, respectively15 (Scheme 5).
O3
Cl1
O2
C4
C12
C5
C3
C1
Cl3
C13
C22
C9
N2
C6
C2
N1
C14
C11
C16
C7
Cl2
C8
C15
O1
C10
N3
Figure 1. ORTEP drawing of compound 6 complexed with CHCl3.
TMSCN
TMSCN
H
O
N
N
H
N
H
O
H
O
O
N
O
TMSCN
O
N
O
N
O
Ph
H
O
H
H
Ph
1a (C1)
4
1a (C1)
3
1a (C2)
TMSCN
TMSCN
O
H
N
O
H
O
O
N
O
N
N
N
H
Ph
N
H
O
H
H
O
O
TMSCN
Ph
H
O
5
1b (C2)
1b (C1)
6
1b (C1)
Scheme 4.

2986
M. Pauvert et al. / Tetrahedron Letters 46 (2005) 2983–2987
O
N
O
N
a
O
H
O
H
a
N
OMe
N
OMe
N
N
83%
96%
CN
(R)-7 (ee>97%)
Scheme 5. Reagents and conditions: (a) Sm(OTf)3, 0.2 equiv, CH2Cl2/MeOH 1:1, 18 h, rt.
O
CN
O
CN
O
CN
O
Ph
Ph
5
6
(S)-7 (ee>99%)
65˚C
65˚C
N
CN
N
CN
N
CN
MeO
O
MeO
O
MeO
O
(S)-8
(R)-8
9
Scheme 6.
If, under similar conditions, the reaction of adduct 3
proceeded as expected to give N-protected 2-cyano-
2H-quinoline (S)-8,15 surprisingly, in light of the above
results, reaction of adduct 4, which was conveniently
carried out at 40 ꢁC for reasons of solubility, proved
to be less satisfying, leading to (R)-8 with a small but
detectable erosion of the chirality. To discover the origin
of this deleterious effect, compound (S)-8 was taken up
in a chloroform–methanol solution and heated at
65 ꢁC. HPLC analyses revealed that the degree of epi-
merisation gradually increased over time to finally deli-
ver a ca. 1:1 mixture of (R)-8 and (S)-8 after 3 days.
Besides, racemic 8 was accompanied with some quantity
of compound 9,16 which strongly suggested that the epi-
merisation process arised via the equilibrium depicted in
Scheme 6. Since adducts 3 and 4 were recovered without
deterioration of their original optical purity, when
heated in the same experimental conditions, it thus
appears that both (S)- and (R)-8 are thermally labile
and prone to easy racemisation, most probably via the
intermediary of compound 9.
References and notes
1. (a) Reissert, A. Chem. Ber. 1905, 38, 1603; (b) Grosheintz,
J. M.; Fischer, H. O. L. J. Am. Chem. Soc. 1941, 63, 2021–
2022.
2. (a) Reviews: McEwen, W. E.; Cobb, R. L. Chem. Rev.
1955, 55, 511–549; (b) Popp, F. D. Adv. Heterocycl. Chem.
1968, 9, 1–25; (c) Popp, F. D. Adv. Heterocycl. Chem.
1979, 24, 187–214.
3. (a) For recent references of asymmetric Reissert-type
processes, see: Diaz, J. L.; Miguel, M.; Lavilla, R. J. Org.
Chem. 2004, 69, 3550–3553; (b) Kuethe, J. T.; Comins, D.
L. J. Org. Chem. 2004, 69, 5219–5231; (c) Comins, D. L.;
Huang, S.; McArdle, C. L.; Ingalls, C. L. Organic Lett.
2001, 3, 469–471; (d) Itoh, T.; Nagata, K.; Miyazaki, M.;
Ohsawa, A. Synlett 1999, 1154–1156.
4. (a) Takamura, M.; Funabashi, K.; Kanai, M.; Shibasaki,
M. J. Am. Chem. Soc. 2000, 122, 6327–6328; (b) Taka-
mura, M.; Funabashi, K.; Kanai, M.; Shibasaki, M. J.
Am. Chem. Soc. 2001, 123, 6801–6808; (c) Ichikawa, E.;
Suzuki, M.; Yabu, K.; Albert, M.; Kanai, M.; Shibasaki,
M. J. Am. Chem. Soc. 2004, 126, 11808–11809.
5. Siek, O.; Schaller, S.; Grimme, S.; Liebscher, J. Synlett
2003, 337–340.
6. Garigipati, R. S.; Sorenson, M. E.; Erhard, K. F.; Adams,
J. L. Tetrahedron Lett. 1993, 34, 5537–5540.
In conclusion, asymmetric synthesis of Reissert com-
pounds 3–6 could be achieved in useful diastereoselectiv-
ities. Chiral auxiliary removal from these adducts
produced N-protected-2-cyano-2H-quinolines (R)-8
and (S)-8 as well as 1-cyano-1H-isoquinolines (R)-7
and (S)-7, that are potentially useful intermediates for
the synthesis of alkaloids.
7. (a) To the best of our knowledge this constitutes the first
observation that exchange of Clꢀ for TfOꢀ can help
improve the diastereoselectivity of the Reissert reaction.
Triflate ion-promoted Reissert reaction has already been
reported, however. For recent exemples, see: Pabel, J.;
Ho¨sl, C. E.; Maurus, M.; Ege, M.; Wanner, K. Th. J. Org.
Chem. 2000, 65, 9272–9275; (b) Yamaguchi, R.; Omoto,
Y.; Miyake, M.; Fujita, K.-I. Chem. Lett. 1998, 547–548;
(c) Yamaguchi, R.; Hatano, B.; Kakayasu, T.; Kozima, S.
Tetrahedron Lett. 1997, 38, 403–406.
Acknowledgements
8. Evain, M.; Pauvert, M.; Collet, S.; Guingant, A. Acta
Cryst. 2002, E58, 1121–1122.
9. X-ray crystallographic analysis for compound 6: A color-
less, rod-like crystal of compound 6, approximately
0.75 · 0.25 · 0.75 mm3 in dimension, was fixed at the tip
of a glass Lindeman capillary by means of a silicon glue.
We are grateful to B. Tanguy and C. Chaumette for
preliminary experiments.
Supplementary data
All measurements were carried out on a Bruker–Nonius
¨
KappaCCD diffractometer with graphite monochromated
MoK-L2,3 radiation. After the absorption correction
(Gaussian integration), the structure was solved by direct
methods (Shelxtl)10 and refined (Jana2000)11 with aniso-
tropic atomic displacement parameters for all non-H
Experimental procedures and analytical data for com-
pounds 3–8 are available. Supplementary data associ-
ated with this article can be found in the online
version, at doi:10.1016/j.tetlet.2005.03.034.

M. Pauvert et al. / Tetrahedron Letters 46 (2005) 2983–2987
2987
atoms. All H atoms were defined with fully restrained
geometry (angles and distances) and riding isotropic
displacement parameters (·1.2). Absolute configuration
was unambiguously determined by refining the Flack
enantiopole parameter using Friedel pairs. The structure
data are as follows: C22H18Cl3N3O3, Mr = 478.8, ortho-
12. (a) Ma, J. C.; Dougherty, D. A. Chem. Rev. 1997, 97,
1303–1324; (b) Yamada, S.; Morita, C. J. Am. Chem. Soc.
2002, 124, 8184–8185.
13. For
a conformational analysis of Reissert adducts
by NMR techniques, see: Pandya, A.; Gibson, H. W.
J. Org. Chem. 1993, 58, 2851–2855, and references cited
therein.
rhombic, P212121, a = 6.19110(10), b = 14.2461(3), c =
3
˚
˚
25.2042(7) A, V = 2222.99(9) A , Z = 4, Dcalcd = 1.430
g · cmꢀ3, F(000) = 984, l(MoK-L2,3) = 0.442 mmꢀ1, T =
293 K, R(obs) = 0.0728, S(obs) = 1.36 (4996 hkl, 281
parameters, cutoff for observed: I/r(I) = 2), FlackÕs
parameter: 0.09(9). Crystallographic details have been
deposited at the Cambridge Crystallographic Data
Center (deposition number CCDC 258634).
14. (a) Evans, D. A.; Coleman, P. J.; Carlos Dias, L. Angew.
Chem. Int., Ed. 1997, 36, 2738–2741; (b) Harrison-Marc-
`
hand, A.; Collet, S.; Guingant, A.; Pradere, J. P.; Toupet,
L. Tetrahedron 2004, 60, 1827–1839.
15. The Rf value of 4(S)-benzyl-oxazolidin-2-one on silica
TLC plates is significantly different from that of com-
pounds 7 and 8 so that the chiral auxiliary could be
recovered and reused, if desired.
10. Sheldrick, G. M. SHELXTL V5.0. Analytical X-ray Instru-
ments Inc., Madison, WI, USA, 1995.
11. Petricek, V.; Dusek, M. Jana2000. Structure Determin-
ation Software Programs. Institute of Physics, Praha,
Czech Republic, 2000.
16. Selected NMR data for compound 9: 1H NMR (300 MHz,
CDCl3): d 3.42 (d, J = 4.8 Hz, 2H, CH2), 3.93 (s, 3H,
OCH3), 6.55 (t, J = 4.8 Hz, 1H, CH–CH2), 7.10–7.28
(m, 3H, H–Ar), 7.65 (d, J = 8.1 Hz, 1H, H–Ar).